The Space Exploration and Analysis Laboratory (SEAL) at Georgia Tech studies how sensor data may be used for spacecraft navigation and space science. We are experts in developing algorithms to extract information from data collected by cameras, telescopes, LIDARs, RF transceivers, and other sensors. Our approach is characterized by a relentless pursuit of algorithmically simple yet theoretically rigorous algorithms that lend themselves to implementation onboard resource-constrained spacecraft.

The SEAL group has active projects from industry and U.S. Government sponsors in the following areas.

Optical Navigation (OPNAV)
Optical Navigation (OPNAV) is the task of spacecraft navigation using images of nearby celestial bodies. A variety of OPNAV techniques exist (e.g., using horizon or landmarks), and the best choice usually depends on (1) the type of body being observed and (2) the distance to the body. The SEAL team at Georgia Tech are pioneers of modern OPNAV methods using both horizon and triangulation methods. Horizon-based methods are popular when close to large celestial bodies (e.g., planets, large moons) with an ellipsoidal global shape. Triangulation-based methods are appropriate when very far away from the observed bodies. Our OPNAV algorithms have been incorporated into numerous high-profile flight missions.

Planetary Terrain Relative Navigation (TRN)
When very close to another celestial body, either in a low orbit or while landing, it is sometimes important to directly estimate the relative state between the spacecraft and the nearby body. This is certainly the case for precision landing, remote sensing, and a host of other scientific and engineering applications. One popular way of achieving this task is through terrain relative navigation (TRN), where direct observations of the surface are used to navigate. These observations can be made with a variety of sensors, such as cameras, altimeters, and LIDARs---and the SEAL research team has experience processing all of these measurement types. Most of our present work is in camera-based TRN. In some situations, it is desirable to achieve a global (or absolute) information about the camera position by associating features in an image with known points in a map. This may be accomplished by matching craters or other surface features. In other situations, a map may not exist (and we don’t wish to build one) and we may perform relative navigation using methods such as visual odometry.

Orbit Determination (OD)
Orbit determination (OD) is one of the classical problems of astrodynamics. Attempts to describe the motion of the planets and other celestial bodies (e.g., comets) date back to antiquity, with modern interpretations dating back hundreds of years (to Kepler, Newton, Gauss, and others). Today, we study orbit determination to recover the orbits of newly discovered celestial bodies (e.g., near earth asteroids) as well as artificial satellites (both working and defunct). These efforts are essential for both planetary protection and space domain awareness (SDA). The SEAL research team has expertise in all forms of OD, but especially in the development of new initial orbit determination (IOD) algorithms for different measurement types or different limitations on the information available. We also study the reformulation of classical IOD problems to improve our ability to generate faithful OD solutions.

In addition to planetary protection and SDA, we also study how IOD onboard a spacecraft may be used to enhance system autonomy (e.g., reinitializing an onboard Kalman filter). We have studied this for a variety of new measurement types, including velocity-only IOD and headings-only IOD.